Antimicrobial Resistance and Genomic Characterization of Campylobacter jejuni and Campylobacter coli Isolated from Retail Chickens in Beijing, China

Abstract

Objective Campylobacter species are the main causes of foodborne illness worldwide, posing significant threats to public health. This study aimed to investigate the antibiotic resistance and genomic characterization of C. jejuni/C.coli from retail chickens in Beijing. Methods Antimicrobial susceptibility testing was conducted on 126 C. jejuni/C. coli isolated from retail chickens in Beijing, following CLSI protocols. Whole genomes of all isolates were sequenced using the Illumina platform. Results More C. coli (83.82%) showed multi-drug resistance than C. jejuni (8.62%). Genomic analysis demonstrated 42 sequence types (STs) and 12 clonal complexes (CCs), from which CC828 and CC52 were dominant. cdtA, cdtB and cdtC encoding cytotoxic protein were present spontaneously in most C. jejuni but not found in any C. coli isolates. The abundances of antibiotic resistance genes (ARGs) and virulence genes (VGs) in C. jejuni and C. coli were significantly different, with ARGs numbered in C. coli and VGs in C. jejuni. Conclusions High prevalence of multi-drug resistance C. coli and C. jejuni isolated from Beijing chickens were challenging clinical antibiotic usages in the treatment of Campylobacter infection. The surveillance of particular C. jejuni and C. coli STs correlated with higher resistance and virulence needs to be strengthened in the future.

1. Introduction

Campylobacter continues to be recognized as one of the major zoonotic pathogens causing foodborne diarrheal illnesses. Approximately 500 million cases of gastrointestinal infections caused by Campylobacter have been reported globally [1]. Currently, 61 species and 16 subspecies belonging to the genus Campylobacter have been identified (http://www.bacterio.net/, accessed on 27 May 2024). Among these species, Campylobacter jejuni and Campylobacter coli altogether are responsible for over 90% of all human Campylobacter gastroenteritis cases, of which approximately 30% of acute enteritis caused by C. jejuni develop severe irritable bowel syndrome with a disease fatality rate about 5/100,000 [1]. Poultry, especially chickens, were the primary reservoirs for Campylobacter spp., mainly C. jejuni and C. coli [2]. If chicken products are contaminated with Campylobacter during food processing or sales, humans may become infected after consuming these contaminated chicken products [3].

Campylobacter infection is a self-limiting process that typically does not require antibiotic treatment [4]. However, for patients suffering severe symptoms from infection or with compromised immune systems, treatment with antibiotics, such as erythromycin, tetracyclines, aminoglycosides and fluoroquinolones, is necessary. Fluoroquinolones, especially ciprofloxacin, have long been considered the main antibiotics for the treatment of Campylobacteriosis [5]. As antimicrobial resistance in Campylobacter becomes more severe in many countries, especially with the emergence of multi-drug resistance (MDR), significant concerns regarding food safety and public health have gained attention internationally [6]. Furthermore, antibiotic resistance genes (ARGs) have been identified in Campylobacter isolates from various sources, especially chicken meats and their related products. Epidemiological research has further confirmed that the prevalence of Campylobacter resistant to clinically relevant antibiotics, such as fluoroquinolones, macrolides and aminoglycoside, has increased remarkably during recent years [7].

Recent studies have identified the major virulence factors involved in the pathogenesis of Campylobacter isolates. Campylobacter can adhere to host cells through the expression of flaA, cadF, jlpA, porA and dnaJ genes, invade intestinal epithelial cells through the expression of ciaB and ceuE genes and produce toxins and survive in host cells through the expression of cdtA, cdtB and cdtC genes. In addition, studies have found a correlation between Campylobacter virulence genes and antibiotic resistance, which indicates a link between antibiotic resistance and the colonization or invasion ability of these bacteria [8].

Since antibiotic resistance and virulence status of Campylobacter isolates could be diversified in various ways depending on the antibiotic usage or investigation period and region, it is necessary to assess the antibiotic resistance and virulence distributing patterns through surveillance. To date, many countries are continuously investigating antibiotic resistance and virulence against pathogenic bacteria at the national level [9]. In 2019, a gastroenteritis outbreak caused by MDR C. coli was identified in Beijing, China, indicating the importance of Campylobacter surveillance to protect human health [10]. In food contamination surveillance in China, C. jejuni and C. coli were mainly isolated from food-producing animals, especially chickens [11]. However, to the best of our knowledge, the genetic traits of the coexistence of antibiotic-resistant genes and virulence genes of Campylobacter in retail chickens in Beijing, China, need to be further elucidated.

The purpose of this study was to investigate the antimicrobial susceptibility and assess the genomic variation of C. jejuni and C. coli isolated from retail chickens in Beijing.

2. Materials and Methods

2.1. Campylobacter Isolates Collection and Preparation

In total, 126 isolates of Campylobacter isolated from retail chickens in Beijing, China, 2018–2020, were used. All isolates were stored at −80 °C in Brucella broth supplemented with 50% glycerol (v/v) and 5% laked sheep blood (v/v). The isolates were revived in Bolton broth, incubated at 42 °C for 48 h under microaerophilic conditions (5% O₂, 10% CO₂ and 85% N₂), streaked on Columbia blood agar plate and incubated at 42 °C for 24 h under microaerophilic conditions (5% O₂, 10% CO₂ and 85% N₂). All isolates were identified by PCR assay [12] and confirmed by Vitek 2 with NH card.

2.2. Antimicrobial Susceptibility Test

The antimicrobial susceptibility of the C. jejuni (n = 58) and C. coli (n = 68) isolates were tested and interpreted using the broth microdilution method, according to the guidelines of the Clinical and Laboratory Standard Institute (CLSI, M45-A3) and EUCAST epidemiological cut-off values (ECOFFs), as shown in Supplementary Table S6. Minimal inhibitory concentrations (MICs) to 6 antimicrobials were determined via broth microdilution, including chloramphenicol (CHL), ciprofloxacin (CIP), erythromycin (ERY), doxycycline (DC), gentamicin (GEN) and tetracycline (TET). Reference strain C. jejuni ATCC 33560 was used as quality control for antimicrobial susceptibility experiments. The isolates that showed resistance to three or more antibiotic categories are defined as multi-drug resistant (MDR) [13,14], while in this study, only antibiotics with CLSI breakpoints were taken into account during the calculation of MDR.

2.3. Whole Genome Sequencing

Genomic DNA (gDNA) was extracted using Omega EZNA Bacterial DNA Kit (Omega Bio-tek, Norcross, GA, USA). The gDNA was sent to Novogene for whole genome sequencing, where Illumina HiSeq protocol was used on an Illumina PE150 platform; SOAPdenovo v2.04 (https://github.com/aquaskyline/SOAPdenovo2, accessed on 27 May 2024), SPAdes v3.15.4 (https://github.com/ablab/spades, accessed on 27 May 2024) and AbySS v2.1.5 (https://github.com/bcgsc/abyss, accessed on 27 May 2024) were used for the initial assembly process. Then, GapCloser v1.12 (https://anaconda.org/bioconda/soapdenovo2-gapcloser, accessed on 27 May 2024) was used to refine the assembly. The species identification was reconfirmed using the SpeciesFinder 2.0 on the Centre for Genomic Epidemiology (CGE) website (https://www.genomicepidemiology.org, accessed on 27 May 2024).

2.4. Screening of Antimicrobial Resistance Genes (ARGs), Virulence Genes and Point Mutations

ARGs and virulence genes of Campylobacter genomes were annotated via Resfinder, Virulence Factor Database (VFDB) databases (updated to 27 May 2024) through ABRicate v1.01 (https://github.com/tseemann/abricate, accessed on 27 May 2024). The thresholds for identity and coverage were all set to 80 percent. Point mutations were predicted by Pointfinder 3.1.0 (https://bitbucket.org/genomicepidemiology/pointfinder, accessed on 27 May 2024) and confirmed by comparing with reference genome NCTC 11168 (GCF_000009085.1).

2.5. Multilocus Sequence Typing

In silico multilocus sequence typing (MLST) was conducted using the MLST database on the public databases for molecular typing and microbial genome diversity (PubMLST) website (https://pubmlst.org/, accessed on 27 May 2024). New alleles and sequence types (STs) were submitted to the PubMLST database. A minimum spanning tree of Campylobacter isolates was generated on the basis of the MLST results using PHYLOVIZ Online (https://online.phyloviz.net/index, accessed on 27 May 2024).

2.6. Core Genome Phylogenetic Analysis

The chromosomes were annotated using the prokaryotic genome annotation tool Prokka (v1.12). Core genomes of all assemblies were determined using Roary (v3.11.0) and aligned using MAFFT (v7.313). A maximum-likelihood phylogenetic tree of the aligned genomes was constructed using FastTree (v2.1.10). The phylogenetic relationship of Campylobacter isolates was visualized using the ChiPlot tool (https://www.chiplot.online/, accessed on 27 May 2024) [15].

2.7. Statistical Analysis

Statistical analysis of the data was performed using R version 4.2.2 (http://www.r-project.org/, accessed on 27 May 2024) and IBM SPSS 26 (IBM SPSS, Armonk, NY, USA). The distribution of Campylobacter was visualized by the Sankey plot. Differences in the total number of ARGs and VFs among different species and sources of Campylobacter were assessed by the Mann–Whitney–Wilcoxon test. Among different Campylobacter, cluster heatmap analyses on the prevalence of antibiotic resistance (AMR), specific ARGs and VFs were performed, respectively.

The association between the resistance profile of each antimicrobial and the presence/absence of ARGs, point mutations and virulence genes was assessed using binary logistic regression models. AMR was considered as a binary dependent variable (0 = non-resistant; 1 = resistant).

The correlation between the resistance profile of each antimicrobial and the presence/absence of ARGs and point mutations was analyzed separately for C. coli and C. jejuni. The ARGs bla_OXA-184, bla_OXA-185 and bla_OXA-465 and point mutations L4 (V82I, T91K, V176I, T177S) and, GyrA (S22G, T86I) were excluded from C. coli correlation analysis for either 100% presence or 100% absence in C. coli. The ARGs aac(6′)-aph(2″), aadE-Cc, erm(B), fexA, tet(L) and bla_OXA-489 and point mutation 23S (A2075G) were excluded from C. jejuni correlation analysis for 100% absence in C. jejuni. The antibiotics ciprofloxacin (100% resistance) was excluded from correlation analysis of C. coli.

The evaluation of the correlation between the resistance profile of each antimicrobial and the presence/absence of virulence genes was also carried out separately for C. coli and C. jejuni. The VFs cadF, pebA, jlpA, porA, kpsD, kpsF, neuA1, wlaN, ciaB, ciaC, flgB, flhB, cdtA, cdtB and cdtC were excluded from C. coli correlation analysis for either 100% presence or 100% absence in C. coli. The VFs cadF, pebA, jlpA, cheA, htrB, kpsD, kpsF, ciaB, ciaC, flgB and flhB genes were excluded from C. jejuni correlation analysis for 100% presence in C. jejuni. The antibiotic ciprofloxacin (100% resistance) was excluded from the correlation analysis of C. coli. The differences between variables were considered statistically significant when p value < 0.05.

3. Results

3.1. Antimicrobial Susceptibility Test

Overall, 124 Campylobacter isolates were resistant to at least one antibiotic. Seventy-four (58.73%) isolates were MDR (Figure 1A). Resistance to ciprofloxacin was detected in C. coli (100%) and C. jejuni (89.66) isolates.

For C. coli isolates, resistance to tetracycline accounted for a large proportion (98.53%), followed by resistance to doxycycline (97.06%), erythromycin (85.29%), gentamicin (77.94%) and chloramphenicol (23.53%). For C. jejuni, similarly, resistance to tetracycline also accounted for a large proportion (62.07%), followed by resistance to doxycycline (60.34%), gentamicin (27.59%), chloramphenicol (12.07%) and erythromycin (8.62%) (Figure 1B).

The resistance rates of C. jejuni isolates against each of the six antibiotics were all lower than those of C. coli isolates. More than half of C. coli (83.82%) and C. jejuni (8.62%) were found to exhibit MDR. Sankey plot showed a widespread prevalence and high diversity of Campylobacter during distribution analysis of sampling locations, sample types and MDR pattern (Figure 1C,D).

3.2. Sequence Types and Clonal Complexes

In total, 42 STs and 13 clonal complexes (CCs) were detected in 126 Campylobacter isolates (Figure 1C and Figure 2).

In C. coli, the most common CC was CC828 C. coli (n = 46; 67.65%), which consisted of 11 STs: ST825, ST830, ST860, ST872, ST1145, ST1586, ST1625, ST3131, ST5511, ST7363 and ST13541. The most frequent CC in C. jejuni was CC52 (n = 11; 18.97%), which comprised ST161 and ST4263. The remaining CCs included relatively fewer STs or a small number of strains. The C. jejuni isolates from chickens (n = 58) showed 28 STs; among these, ST161 (n = 7; 12.07%) and ST6681 (n = 7; 12.07%) were the most prevalent. The C. coli isolates from chickens (n = 68) showed 14 STs, among which ST6322 (n = 20; 29.41%), ST1625 (n = 9; 13.24%) and ST3131 (n = 8; 11.76%) were the most common. In addition, after verification by PubMLST, a new ST13541 of C. coli was identified in this study. No C. coli isolates from Clade 2 or 3 were found in this study.

3.3. Analysis of Antibiotic Resistance Genes and Point Mutations

A total of 15 different ARGs and 10 point mutations encoding resistance to six antimicrobial classes were detected in 126 isolates, including β-lactam (n = 5), aminoglycoside (n = 6), quinolone (n = 2), phenicol (n = 2), tetracycline (n = 2) and macrolide (n = 8) (Figure 3). The gene encoding β-lactam resistance was observed for bla_OXA-193 (84.13%), followed by tet(O) (80.95%) encoding tetracycline resistance. Detection rates of the remaining ARGs were <50%.

High prevalence of mutation T86I in protein GyrA was found in 96.03% (121/126) of all the Campylobacter isolates, while S22G was found in only 36.51% (46/126) of them. One mutation of RpsL (K43R) and six mutations (V82I, T91K, V121A, V176I, T177S and M192I) on ribosomal protein L4 was found, from which V121A and M192I were present in 31 and 16 isolates. Mutation on the 23S rRna gene (A2075G) was found in 16 isolates.

In C. coli, high prevalence rates of GyrA (T86I) (100%), bla_OXA-193 (86.76%) encoding β-lactam resistance and tet(O) (86.76%) and aac(6′)-aph(2″) (75%), ant(6)-Ia (73.53%) and aph(3′)-III (58.82%) encoding aminoglycoside resistance were observed.

In C. jejuni, the most frequent one was GyrA (T86I) (91.38%), followed by bla_OXA-193 (81.03%) encoding β-lactam resistance and GyrA (S22G) (79.31%) and tet(O) (74.14%) encoding tetracycline resistance (Figure 3). The average number of ARGs in C. coli was significantly higher than that in C. jejuni (Figure 3). The genes aac(6′)-aph(2″) (75%), erm(B) (41.18%), 23S (A2075G) (23.53%), aadE-Cc (4.41%), fexA (2.94%), tet(L) (2.94%) and bla_OXA-489 (1.47%) were detected only in C. coli, while GyrA (S22G) (79.31%), L4 V82I (6.89%), bla_OXA-184 (6.90%), bla_OXA-185 (6.90%), bla_OXA-465 (3.45%), L4 T177S (1.72%), L4 V176I (1.72%) and L4 T91K (1.72%) were detected only in C. jejuni (Figure 3).

3.4. Analysis of Virulence Genes

In total, 21 virulence genes classified into five VF classes, such as adherence, immune modulation, invasion, motility and toxin, were identified in 126 Campylobacter isolates (Figure 4). High prevalence rates were observed for the virulence genes cadF (100%) and pebA (100 %) encoding adherence; kpsD (100%), kpsF (100%), cheA (99.21%) and kpsM (56.35%) encoding immune modulation; ciaB (100%) and ciaC (100%) encoding invasion; and flgB (100%) and flhB (100%) encoding motility (Figure 4). Detection rates of the remaining virulence genes were less than 50 %. Type VI secretion system was absent in all isolates.

This study also investigated the effects of various factors on the number of virulence genes carried by the Campylobacter isolates. Similar to ARGs, the mean number of virulence genes in Campylobacter was found to be significantly higher in isolates from supermarkets than those from farmers’ markets (p < 0.05), whereas the mean number of virulence genes in Campylobacter isolates from black-bone chickens was significantly higher than that in isolates from local chickens.

All C. coli and C. jejuni harbored the virulence genes cadF, pebA, kpsD, kpsF, ciaB, ciaC, flgB and flhB correlated with adherence, immune modulation, invasion and motility functions. The average number of virulence genes in C. jejuni was significantly higher than that in C. coli (Figure 5C). Notably, C. jejuni harbored all three cdt genes at high prevalence levels, such as cdtA (93.10%), cdtB (98.27%) and cdtC (98.27%), while cdtA, cdtB and cdtC genes were not found in all C. coli isolates (Figure 4).

3.5. Phylogenetic Analysis

To investigate the genetic relationship among Campylobacter isolates from retail chickens in Beijing, a maximum likelihood tree was constructed using core single nucleotide polymorphisms (SNPs) identified in the 126 Campylobacter isolates (Figure 4). As shown in the phylogenetic tree, the 126 Campylobacter strains clustered distinctly into two branches: C. coli and C. jejuni.

The 68 C. coli isolates belonged to CC828, CC1150 and an unassigned CC, whereas the 58 C. jejuni isolates belonged to 10 distinct CCs (CC21, CC48, CC49, CC52, CC57, CC607, CC353, CC354, CC443 and CC46) and an unassigned CC.

In C. coli, seven isolates from ST830 and ST872 displayed the least resistance genes and were sensitive to most antimicrobial agents. Seventeen isolates from ST860, ST1145, ST6332 and ST7363 harbored the most resistance genes and were resistant to most antimicrobial agents. Although ARGs were comparatively less prevalent in C. jejuni than in C. coli, it is intriguing that seven isolates of ST653, ST6717, ST6681, ST6683 and ST7360 carry more resistance genes than other C. coli (Figure 3 and Figure 4).

3.6. Association between Antibiotic Resistance and Presence of ARGs and Point Mutations

Binary logistic regression models were applied for each Campylobacter isolate, using the antimicrobial resistance profiles as a dependent variable and the presence/absence of ARGs and point mutations as independent variables. It was observed that C. coli and C. jejuni showed a statistically significant association (p < 0.05) between antimicrobial resistance and the presence of ARGs. In C. coli isolates, the tet(O) gene showed a positive association with chloramphenicol (Figure 5 and Supplementary Table S1). In C. jejuni isolates, the aph(2″)-If gene showed a positive association with erythromycin resistance, the tet(O) gene showed a positive association with erythromycin resistance, and the L4 M192I gene showed a positive association with gentamicin and chloramphenicol resistance (Figure 5 and Supplementary Table S2).

3.7. Association between Antibiotic Resistance and Presence of Virulence Genes

Binary logistic regression models were applied for each Campylobacter isolate, using the antimicrobial resistance profiles as a dependent variable and the presence/absence of virulence genes as an independent variable. It was observed that C. coli and C. jejuni showed a statistically significant association (p < 0.05) between antimicrobial resistance and the presence of virulence genes. In C. coli isolates, the Cj1135 gene showed a negative association with resistance to doxycycline; the kpsM gene showed a negative association with resistance to gentamicin, erythromycin and chloramphenicol (Figure 5 and Supplementary Table S3). In C. jejuni isolates, the Cj1135 gene showed a positive association with resistance to gentamicin, erythromycin, ciprofloxacin and tetracycline (Figure 5 and Supplementary Table S4).

4. Discussion

Campylobacter, one of the four main causes of gastroenteritis worldwide, poses a serious threat to food safety and public health. At present, the contamination of MDR Campylobacter isolates in food products and their hypervirulent potential have globally raised public safety concerns. However, there is limited research in China on the antibiotic resistance phenotypes, ARGs and virulence genes of Campylobacter isolated from chickens sold in supermarkets and farmers’ markets. In this study, we first evaluated the antibiotic resistance phenotype and MDR of Campylobacter in retail chickens from supermarkets and farmers’ markets in Beijing and revealed the molecular characteristics of Campylobacter isolates. Genomic analysis of pathogens in retail chickens provides a more comprehensive and better understanding of the characteristics of Campylobacter isolated from chickens, which may be used to inform prevention strategies. Nevertheless, few studies have been conducted in China to examine the genomic characteristics of Campylobacter in retail chickens from supermarkets. The genome sequences of Campylobacter obtained in this study can provide a source of information on pathogens in retail chickens for researchers worldwide.

Owing to the extensive use of antibiotics in human health care, livestock production and agriculture, antibiotic resistance has emerged as a major global concern [16]. Of all the 126 Campylobacter isolates in our study, 49.21% (62/126) were MDR, and 15.08% (19/126) were resistant to antibiotics of at least five antimicrobial classes. The highest resistance to ciprofloxacin, tetracycline and doxycycline was observed in this study, which was consistent with previous reports [17]. The proportion of MDR isolates in this study is markedly higher than that reported in Japan, Spain and Canada [18,19,20]. The resistance of C. coli was more severe than that of C. jejuni, especially ST6322 C. coli isolates, which were 95% (19/20) MDR. However, it is worth noting that the antibiotic resistance rate of C. jejuni isolated from diarrhea patients in Beijing is generally lower than that of this study [21]. Remarkably, four C. jejuni isolates showed resistance to six antibiotics, which indicates that strengthened surveillance of the AMR status in Campylobacter needs to be implemented in the future.

For ARGs, we found that a high percentage of C. coli (32.35%) and C. jejuni (77.59%) strains carried tet(O), which is consistent with previous reports, indicating that Campylobacter exhibits high levels of tetracycline resistance [22]. According to reports, the tet(O) gene is the only tetracycline resistance determinant found in Campylobacter and is widely detected in all tetracycline-resistant Campylobacter isolates [22]. The mutation T86I in protein GyrA was associated with quinolone resistance in Campylobacter [23]. All isolates showed resistance to ciprofloxacin, and the high prevalence of GyrA (T86I) (96.03%, 121/126) in all Campylobacter isolates explained resistance to ciprofloxacin in 121 isolates. The other five isolates resistant to ciprofloxacin without T86I were found. Another mutation (S22G) out of the QRDR of GyrA, which correlated with quinolone resistance [24], was found in another five isolates resistant to ciprofloxacin but without T86I mutation. The high prevalence of the tet(O) gene in our isolates indicates that the Campylobacter isolates from Beijing chickens have high tetracycline resistance. Among aminoglycoside resistance genes, aadE-Cc, aph(3′)-III, aac(6′)-aph(2″) and ant(6)-Ia were observed to be more common in C. coli than in C. jejuni in the present study, correlating well with resistance phenotypes of C. coli and C. jejuni, which is consistent with previous findings [22]. The cat gene and fexA gene encode resistance to phenicols, whereas the erm(B) gene encodes resistance to macrolides. Remarkably, although the cat (10.34%) gene was detected at a lower frequency in C. jejuni, and the fexA gene and erm(B) gene were not detected, C. jejuni still exhibited resistance to chloramphenicol (12.07%) and erythromycin (8.62%). It has been reported that mutations on RpsL, ribosomal protein L4 and 23S rRNA gene were responsible for macrolides resistance of Campylobacter [25]. In our study, we discovered RpsL (K43R) and six mutations (V82I, T91K, V121A, V176I, T177S and M192I) on ribosomal protein L4, and the 23S rRNA gene (A2075G) was found mostly in cat-, erm(B)- and fexA-absent C. jejuni isolates, partially implicating the mechanism of macrolides resistance of C. jejuni.

In this study, MLST typing was used to investigate the genotypes of chicken-derived Campylobacter from supermarkets and farmers’ markets in Beijing. Our findings demonstrated that 126 Campylobacter isolates were classified into 42 STs, with strains belonging to CC828 as the predominant group. CC828 is the largest and most widely distributed CC worldwide, representing 21.08% of all Campylobacter strains submitted to the PubMLST database. As reported by numerous studies, the major CCs of Campylobacter vary by country and region, but CC21 (16.39%), CC353 (7.10%), CC45 (6.12%) and CC48 (5.01%) are consistently the predominant CCs among isolates in the PubMLST database. CC828 is the most frequently isolated CC from chickens, and the majority of isolates belonging to CC828 are C. coli isolates [26]. In the present study, the prevalences of CC52 (n = 11; 8.73%) and CC353 (n = 9; 7.14%) were lower than that of CC828. It is worth noting that the main source of CC52 isolates is humans, but the prevalence of CC52 isolates in chickens was relatively high in this study. This result is consistent with previous findings, suggesting a potential bidirectional transmission of these strains between humans and chickens [27]. Notably, genetic evolution analysis showed that not all isolates with the same STs originated from the same sampling location and type of chicken, and some ST isolates were cross-distributed in different supermarkets and types of chicken; this finding is consistent with previously reported findings [28]. Campylobacter isolated from broiler chickens in eastern China mainly includes C. jejuni ST8089, ST10242, ST10244 and ST10243 and C. coli ST1121, ST830, ST1568, ST1625, ST872 and ST829 [29]. Maesaar et al. (2018) reported ST5, ST45 and ST50 as the most prevalent genotypes of C. jejuni isolates from poultry chickens in Estonia [30]. The prevalent STs of C. jejuni isolated in the present study do not correspond with those mentioned above. However, the detected C. coli STs ST830, ST872, ST1121, ST1568 and ST1625 are almost identical to the ones mentioned above. These data highlight the wide-range transmission of these strains across regions and indicate that ST diversity varies among countries and regions. In addition, a C. coli strain (BJWQ2) belonged to an ST that had not been reported previously. Sequence data from this strain were submitted to the Campylobacter MSLT database (PubMLST), which led to the assignation of a novel ST (ID PubMLST 13541) and a novel allele sequence for gltA (ID PubMLST 841).

The pathogenesis of Campylobacter to cause diarrhea is complicated. Previous studies showed that genes involved in epithelial cell motility, colonization, invasion and toxin production play an important role in the development of Campylobacter-related diseases [31]. The flaC, ciaB, pebA and cadF genes were highly prevalent in the Campylobacter isolates in this study; nevertheless, there are divergences regarding the constitutive presence of these genes in this microorganism [31]. The presence of the cadF gene is crucial for pathogenesis, as adhesion is a prerequisite process for any bacterial pathogens to invade epithelial cells; our results demonstrated ubiquitous invading potentiality in all Campylobacter isolates from retail chickens in this study. The invasion of epithelial cells and the production of CDT are important bacterial virulence mechanisms that play a key role in enterocolitis. While the presence of a single cdt gene has no effect on bacterial virulence, the coexistence of all three cdt genes leads to the production of functional cytotoxic substances [32]. In our study, cdtA, cdtB and cdtC were present spontaneously in most C. jejuni isolates but were not found in all C. coli isolates, indicating that C. jejuni is more virulent than C. coli in retail chickens in Beijing.

Phylogenetic analysis provides information on the prevalence and phylogenetic relationships of antibiotic resistance phenotypes, ARGs and virulence genes of Campylobacter. A previous study found that the virulence genes cadF and ciaB affect chloramphenicol and ampicillin resistance of Campylobacter [8]. Similarly, in our study, analysis of the relationship between the presence of virulence genes and antimicrobial resistance in C. jejuni isolates showed consistent positive (p < 0.05 and an OR > 1) correlations. The mechanisms underlying the correlation between antibiotic resistance and virulence genes in Campylobacter remains unclear to date, for which further research is needed.

5. Conclusions

Antimicrobial resistance of C. coli from retail chickens in Beijing is severe, especially with the high prevalence of MDR C. coli isolates. Although the antimicrobial resistance level of C. jejuni isolates is relatively milder than C. coli, their stronger invading potentiality is of great concern to public health. Therefore, it might be challenging for clinical antibiotic usage in the treatment of illnesses caused by Campylobacter. Meanwhile, the distinctive sequence types of C. jejuni and C. coli correlating with higher resistance and virulence genotypes in chickens need to be considered, particularly in future surveillance.